Aligning Sentence Embeddings to Human Concepts via Sparse Autoencoders

• URL: http://arxiv.org/abs/2607.00023v1
• Date: Fri, 19 Jun 2026 02:39:50 GMT
• Title: Aligning Sentence Embeddings to Human Concepts via Sparse Autoencoders
• Abstract summary: We propose a method to disentangle the dense representations of sentence transformers (e.g., E5) into human-interpretable concepts using Top-k Sparse Autoencoders (SAEs)<n>We demonstrate that these disentangled features align with specific semantic, syntactic, and pragmatic categories.<n>Our findings suggest that SAE-based decomposition offers a viable path toward transparent and steerable neural information retrieval.
• Score: 5.053354418366577
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Dense sentence embeddings are fundamental to modern Retrieval-Augmented Generation (RAG) systems but suffer from a lack of interpretability due to feature superposition. This opacity hinders the alignment of retrieval processes with human intent, as the entangled representations are difficult to analyze or control. In this work, we propose a method to disentangle the dense representations of sentence transformers (e.g., E5) into human-interpretable concepts using Top-k Sparse Autoencoders (SAEs). We demonstrate that these disentangled features align with specific semantic, syntactic, and pragmatic categories. Furthermore, we introduce an activation steering mechanism that allows for precise intervention in the retrieval process. By clamping specific latent features, we show that it is possible to re-rank search results to better align with user constraints without retraining the backbone model. Our findings suggest that SAE-based decomposition offers a viable path toward transparent and steerable neural information retrieval.

Related papers

• Decoupling Semantics and Logic: A Training-Free Coarse-to-Fine Pipeline for Video Retrieval-Augmented Generation [5.263138429758972]
  This paper presents our system description for the 2nd Workshop on Multimodal Augmented Generation via MultimodAl Retrieval (MAGMaR)<n> Addressing the critical challenges of cross-lingual long-video comprehension, strict persona adherence, and zero-hallucination temporal grounding, we propose a two-stage cascaded Video RAG pipeline.<n>Our architecture strategically decouples semantic retrieval from cognitive logical reasoning through a modality-aware division of labor.
  arXiv  Detail & Related papers  (2026-06-06T01:17:40Z)
• Reward-Guided Semantic Evolution for Test-time Adaptive Object Detection [82.2968697030677]
  Open-vocabulary object detection with vision-language models (VLMs) such as Grounding DINO suffers from performance degradation under test-time distribution shifts.<n>We propose Reward-Guided Semantic Evolution (RGSE), a training-free framework that directly refines the text embeddings at test time.
  arXiv  Detail & Related papers  (2026-05-06T06:17:41Z)
• Beyond Semantic Similarity: Rethinking Retrieval for Agentic Search via Direct Corpus Interaction [127.64173950476702]
  We study direct corpus interaction (DCI), where an agent searches the raw corpus directly with general-purpose terminal tools.<n>This approach requires no offline indexing and adapts naturally to evolving local corpora.<n>Across IR benchmarks and end-to-end agentic search tasks, this simple setup substantially outperforms strong sparse, dense, and reranking baselines.
  arXiv  Detail & Related papers  (2026-05-03T19:13:11Z)
• AlignSAE: Concept-Aligned Sparse Autoencoders [47.18866175760984]
  We introduce AlignSAE, a method that aligns SAE features with a defined ontology through a "pre-train, then post-train" curriculum.<n>After an initial unsupervised training phase, we apply supervised post-training to bind specific concepts to dedicated latent slots.<n>This separation creates an interpretable interface where specific relations can be inspected and controlled without interference from unrelated features.
  arXiv  Detail & Related papers  (2025-12-01T18:58:22Z)
• Interpreting CFD Surrogates through Sparse Autoencoders [6.842974489069953]
  This work introduces a posthoc interpretability framework for graph-based surrogate models used in computational fluid dynamics (CFD)<n>By obtaining an overcomplete basis in the node embedding space of a pretrained surrogate, the method extracts a dictionary of interpretable latent features.<n>The approach enables the identification of monosemantic concepts aligned with physical phenomena such as vorticity or flow structures.
  arXiv  Detail & Related papers  (2025-07-21T21:09:45Z)
• Decoding Dense Embeddings: Sparse Autoencoders for Interpreting and Discretizing Dense Retrieval [15.665567982431924]
  We propose a novel interpretability framework for Dense Passage Retrieval (DPR) models.<n>We generate natural language descriptions for each latent concept, enabling human interpretations of both the dense embeddings and the query-document similarity scores of DPR models.<n>We show that Concept-Level Sparse Retrieval (CL-SR) achieves high index-space and computational efficiency while maintaining robust performance across vocabulary and semantic mismatches.
  arXiv  Detail & Related papers  (2025-05-28T02:50:17Z)
• Disentangling Dense Embeddings with Sparse Autoencoders [0.0]
  Sparse autoencoders (SAEs) have shown promise in extracting interpretable features from complex neural networks. We present one of the first applications of SAEs to dense text embeddings from large language models. We show that the resulting sparse representations maintain semantic fidelity while offering interpretability.
  arXiv  Detail & Related papers  (2024-08-01T15:46:22Z)
• Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models [55.19497659895122]
  We introduce methods for discovering and applying sparse feature circuits.<n>These are causally implicatedworks of human-interpretable features for explaining language model behaviors.
  arXiv  Detail & Related papers  (2024-03-28T17:56:07Z)
• Tuning-Free Accountable Intervention for LLM Deployment -- A Metacognitive Approach [55.613461060997004]
  Large Language Models (LLMs) have catalyzed transformative advances across a spectrum of natural language processing tasks. We propose an innovative textitmetacognitive approach, dubbed textbfCLEAR, to equip LLMs with capabilities for self-aware error identification and correction.
  arXiv  Detail & Related papers  (2024-03-08T19:18:53Z)
• Adversarial Semantic Data Augmentation for Human Pose Estimation [96.75411357541438]
  We propose Semantic Data Augmentation (SDA), a method that augments images by pasting segmented body parts with various semantic granularity. We also propose Adversarial Semantic Data Augmentation (ASDA), which exploits a generative network to dynamiclly predict tailored pasting configuration. State-of-the-art results are achieved on challenging benchmarks.
  arXiv  Detail & Related papers  (2020-08-03T07:56:04Z)
• Closed-Form Factorization of Latent Semantics in GANs [65.42778970898534]
  A rich set of interpretable dimensions has been shown to emerge in the latent space of the Generative Adversarial Networks (GANs) trained for synthesizing images. In this work, we examine the internal representation learned by GANs to reveal the underlying variation factors in an unsupervised manner. We propose a closed-form factorization algorithm for latent semantic discovery by directly decomposing the pre-trained weights.
  arXiv  Detail & Related papers  (2020-07-13T18:05:36Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.